I will have a go at answering my question, although I am not certain I have it right. I think the expected RMS displacement between two points undergoing independent random walks in dimensions after n steps is given by
with a standard deviation of
This seems to agree with the outcome of simulations, but I do not have a good source to back it up.